Keywords: Analysis/Processing, Data Processing

Motivation: Pre-processing MR images is a necessary step prior to image analysis due to variability of intensity scale in MR images.

Goal(s): To develop a deep learning algorithm for standardizing knee MR images prior to analysis.

Approach: We developed a denoising autoencoder with VNet architecture achieving on-the-fly image pre-processing (Bias field correction and intensity normalization) and denoising. Image quality was evaluated using SNR, NMSE, PSNR, and SSIM.

Results: Our approach achieved an improved SNR with an efficient runtime compared to conventional pre-processing methods.

Impact: Our DL-based knee MRI pre-processing tool generates standardized MRI outputs for image analysis and DL model development. This tool can be incorporated into a wide range of image analysis pipelines for the knee.

Keywords: AI/ML Image Reconstruction, Machine Learning/Artificial Intelligence

Motivation: 3D medical images are often acquired with anisotropic volumes to reduce scan times. Super-resolution reconstruction to recover features in the low-resolution direction would improve visualisation and clinical accuracy.

Goal(s): To train an unpaired super-resolution network for anisotropic 3D MRI and CT images.

Approach: We propose that it is possible to leverage disjoint patches from the high-resolution (in-plane) data to increase the resolution of the low-resolution (through-plane) slices.

Results: We demonstrate that our proposed modified CycleGAN architecture, performs better than the standard CycleGAN for super-resolution of MRI and CT data.

Impact: Unpaired super-resolution reconstruction of anisotropic 3D medical images, enables accurate recovery of features in the low-resolution direction of MRI and CT data.

Keywords: AI/ML Image Reconstruction, Machine Learning/Artificial Intelligence

Motivation: Multi-contrast MRI with a single fully-sampled HR image and multiple under-sampled, LR contrasts can streamline diagnostics while reducing scan times.

Goal(s): To effectively reconstruct under-sampled images by extracting anatomical information from the fully-sampled HR reference.

Approach: Employ contrast/anatomy disentanglement learning to preserve anatomical consistency and restore unique contrast features in the LR images.

Results: Preliminary outcomes indicate superior reconstruction fidelity compared to traditional methods, enhancing diagnostic quality and efficiency.

Impact: By preserving imaging quality while reducing MRI scan times, patient convenience is substantially enhanced, allowing scalability across various contrasts.

Keywords: AI/ML Image Reconstruction, Image Reconstruction, deep learning; internal auditory canal; high resolution T2-weighted imaging; super-resolution.

Motivation: There are no reports on the usefulness of high-resolution three-dimensional T2-weighted imaging (HR-3D T2WI) using super-resolution deep learning-based reconstruction (SR-DLR) technique.

Goal(s): We aimed to determine whether a novel SR-DLR technique can improve the image quality of HR-3D T2WI for assessing the IAC and inner ear.

Approach: We qualitatively assessed the image quality of HR-3D T2WI with and without SR-DLR for the visualization of detailed structures of the IAC and inner ear.

Results: SR-DLR was shown to significantly improve the image quality of HR-3D T2WI in visualizing nerves in the IAC and inner ear compared to conventional HR-3D T2WI without SR-DLR.

Impact: High-resolution three-dimensional T2-weighted imaging with super-resolution deep learning-based reconstruction can improve the visualization of detailed structures of the internal auditory canal and inner ear without extended acquisition time.

Keywords: AI/ML Image Reconstruction, Machine Learning/Artificial Intelligence, pituitary microadenoma

Motivation: The spatial resolution of MRI is still limited to the detection of pituitary microadenomas. CSAISR framework can reduce noise and improve image resolution.

Goal(s): To assess the image quality and pituitary microadenoma detection performance of the thin-slice MRI using CSAISR framework.

Approach: In this work, 1.5mm-CSAISR, 1.5mm-CSAI, 1.5mm-CS and 3mm-CSAISR images were obtained.  These 1.5mm images were evaluated subjectively and objectively, and the detection rate of 1.5mm-CSAISR and 3mm-CSAISR were compared.

Results: Combined with subjective and objective evaluation, the image quality of 1.5mm-CSAISR images was the best. Meanwhile, the detection rate of 1.5mm-CSAISR reached 92.5%, which was significantly better than that of 3mm-CSAISR.

Impact: The results suggest that thin-slice MRI combined with CSAISR framework can balance the relationship between noise reduction and spatial resolution improvement,  increase the detection rate of pituitary microadenoma and is meaningful for the diagnosis, follow-up and localization of this disease.

Keywords: AI/ML Image Reconstruction, Brain Connectivity, Analysis/Processing, AI/ML Image Reconstruction

Motivation: Compressed Sense (CS) acquisition in combination with novel deep learning-based reconstructions has been shown as a viable acceleration technique that brings about additional artificial intelligence (AI) based denoising.

Goal(s): Here, we investigate the impact of CS-AI acceleration and denoising on high-resolution resting-state (rs)-fMRI analysis.

Approach: CS was performed, and different reconstruction methods were compared: (i) conventional CS, (ii) CS with moderate SmartSpeed AI based denoising and (iii) CS with strong SmartSpeed AI based denoising.

Results: Our preliminary results indicate that the underlying reconstruction CS nets do not introduce “artificial” noise or bias and are capable of generating the expected neuronal networks.

Impact: Increasing the rs-fMRI resolution, without sacrificing fidelity in functional connectivity maps, via the application of CS-AI, will lead to higher confidence in human brain mapping, thus reducing the number of participants needed to detect differences between healthy and clinical populations.

Keywords: AI/ML Image Reconstruction, Brain, Super-Resolution

Motivation: The motivation behind this study is to alleviate discomfort during clinical exams by improving through-plane resolution.

Goal(s): Our objective is to develop a deep learning-based super-resolution approach for low-resolution T2 images, reducing patient discomfort and improving diagnosis accuracy.

Approach: We develop a deep learning framework that combines information from T1 and T2 scans, enabling the generation of high-quality images in the through-plane direction.

Results: The proposed approach successfully enhances through-plane super-resolution in brian MRI, resulting in superior image quality. This improvement has the potential to improve diagnostic accuracy and alleviate patient discomfort during clinical exams.

Impact: This study presents a novel deep learning framework that improves through-plane Super-Resolution in brain MRIs, thus enhancing diagnostic accuracy and reducing patient discomfort during routine health checks.

Keywords: Data Processing, Brain

Motivation: This study investigates the impact of deep learning-based image reconstruction (DLR) in structural brain MRI volumetric analysis.

Goal(s): To demonstrate that DLR effectively reduces noise and enhances image quality with short acquisition time, 

Approach: Ten healthy subjects were scanned with a 3T MRI system with and without DLR reconstruction.

Results: Voxel-based morphometry analysis revealed significant improvements in brain volumetric measurements with DLR compared to conventional methods. These advancements are particularly relevant in regions associated with neurodegenerative diseases.

Impact: DLR offers the potential to facilitate earlier detection and monitoring of such conditions, providing clinical value with comparable scan duration.

Keywords: Analysis/Processing, DSC & DCE Perfusion

Motivation: Dynamic contrast-enhanced MRI (DCE-MRI) is invaluable for non-invasive assessment of tissue perfusion and microcirculation dynamics. However, unreliability of DCE-MRI discourages clinical application.

Goal(s): To evaluate the image quality and diagnostic performance of enhanced DCE-MRI using a deep learning-based super-resolution and denoising algorithm.

Approach: Deep learning-based super-resolution and denoising (DLSD) algorithm was applied to DCE-MRI obtained from 306 patients with adult-type diffuse gliomas to reduce noise and increase resolution.

Results: DLSD significantly enhanced image quality without compromising diagnostic accuracy in distinguishing low- and high-grade tumors and IDH mutation, and it also improved the reliability of arterial input functions.

Impact: Improving DCE-MRI image quality and reliability through deep learning-based super-resolution and denoising algorithm can help address previous reliability issues and offer clinical applicability not only in the field of diffuse glioma but also in other areas utilizing DCE-MRI.

Keywords: Analysis/Processing, CEST & MT

Motivation:  B₁-correction is important in CEST MRI. Conventional correction methods require multiple B₁ acquisitions, making them less practical in clinical adoptions.

Goal(s): This study proposed a CNN-based model that can correct B₁ inhomogeneity from a single B₁ acquisition.

Approach: A CECNN-B₁ model incorporating B₁ distribution was designed to enhance the confidence of CEST quantification, and its performance was evaluated on a creatine phantom at 5 T.

Results: Substantial variation of B₁ distribution was observed, resulting in inhomogeneous CEST map. After correction, the B₁-induced artifact was effectively alleviated. The image quality of the corrected CEST map was superior to that calculated with conventional interpolation method.

Impact: The proposed CECNN-B₁ model enabled B₁ inhomogeneity correction from single B₁ acquisition. A creatine phantom study showed its superiority over the conventional interpolation methods requiring multiple B₁ acquisitions, providing an efficient way for improved CEST MRI on high-field scanners.

Keywords: Analysis/Processing, Quantitative Susceptibility mapping

Motivation:
Lengthy acquisitions are needed to produce high-quality quantitative susceptibility mapping (QSM) from which it is possible to segment vasculature and extract physiological parameters.

Goal(s): To adapt a deep learning method for super-resolution reconstruction to enhance QSM images.

Approach: We applied the 3D densely-connected super resolution network (DCSRN) to QSM data, as it has previously shown promising results in reconstructing T1w high-resolution (HR) images from low-resolution (LR) images.

Results: We demonstrated an improvement in the reconstruction of the vascular network, with intravascular susceptibility values distribution close to the true distribution.
 

Impact: Our results show the promise of DCSRN architecture in producing super resolution (SR) images from low resolution (LR) images. Furthermore, the feasibility of segmenting vessels and extracting venous OEF on SR would be beneficial for studies of brain vasculature.

Keywords: Analysis/Processing, Data Analysis

Motivation: When quantifying MRI image quality, image similarity metrics must be able to detect image artifacts.

Goal(s): To investigate the sensitivity of image similarity metrics to common distortion sources in MRI.

Approach: Distorted MRI is simulated using blurring, noise, inter-scan and intra-scan motion. Regression forests are trained to estimate the distortion parameters based on the image similarity metrics. The regression forests' feature importance quantifies the image metric sensitivity.

Results: Not all image similarity metrics are equally sensitive to every distortion source, and the best metric depends on the distortion source. The appropriate metric must be used to quantify the image quality.

Impact: Typically, standard image similarity metrics such as SSIM are chosen to estimate whether a particular method outperforms another method for all tasks. However, this research can help scientists use the appropriate metric when evaluating MRI reconstruction and processing methods.

Keywords: Analysis/Processing, PET/MR, Low Dose PET/MR

Motivation: We developed a deep learning model to enhance the image quality of ultra-low dose brain PET.

Goal(s): Significantly reducing the injected dose not only minimizes radiation risk in subjects but also provides options for scanning protocols, and more follow-up studies.

Approach: We proposed a 3D-Residual Attention U-Net model initially trained on whole-body [18F]FDG PET/MR images. We used transfer learning approach to fine-tune our proposed model on [18F]CPFPX PET/MRI inhouse dataset.

Results: We achieved improved metrics compared to U-Net model with average PSNR of 28.02 (U-Net: 21.23), SSIM of 0.81 (U-Net: 0.53), CNR of 0.72 (U-Net: 0.61) and NMSE of 0.33 (U-Net: 0.67).

Impact: Our model has potential to generate high-quality PET images from low-dose PET/MR, potentially contribute to implementation of kinetic modelling using PET/MR imaging. Our model is capable of enhancing both whole-body and brain datasets, making it valuable asset for diverse applications.

Keywords: Analysis/Processing, Velocity & Flow, Velocity enhancement, Anti-aliasing

Motivation: 4D flow MRI suffers from different sources of noise and aliasing artifacts. However, the existing techniques for enhancing velocities in 4D flow MRI encounter reliability problems with varying flow patterns or acquisition parameters.

Goal(s): Our goal was to develop a velocity enhancement an anti-aliasing technique for 4D Flow MRI that can be easily applied to diverse flow types.

Approach: We incorporate a vector potential into a neural network to predict velocity fields that strictly adhere to the divergence-free condition.

Results: Results from simulated 4D flow MRI images demonstrate significant noise reduction and aliasing correction.

Impact: The proposed Physics-Informed Neural Network enables the recovery of noise-free and aliasing artifact-free velocity fields using divergence-free terms in the network without the need for tuning hyperparameters in the training function, enhancing the applicability of these networks to different datasets.

Keywords: Analysis/Processing, Machine Learning/Artificial Intelligence, Angiography

Motivation: CE MRA data is susceptible to motion and noise artifact due to its longer acquisition time. Conventional intensity-based registration are often unreliable, necessitating better artifact correction methods.

Goal(s): To provide better artifact correction methods for CE MRA using generative deep learning and angiogram-aware loss function

Approach: two deep learning architectures were trained with/without angiogram-aware loss function. Network accuracy was evaluated based on CE MRA dynamic scans and angiogram.

Results: motion correction was successfully performed, resulting in angiograms with PSNR=37.9±4.3 and SSIM=0.97±0.04. angiogram-aware loss function improved the correction accuracy by up to 13 points in PSNR and 17 points in SSIM.

Impact: We developed accurate deep learning solutions for CE MRA artifact correction, potentially reducing the need for repeated MRA scans. We also showed that angiogram-aware loss function, which considers the last processing steps of CE MRA data, can improve correction accuracy.